LINGUIST List 2.140

Tuesday, 16 Apr 1991

Disc: (Morpho)phonology

Editor for this issue: <>

Directory

1. , (Morpho)phonology
2. , Re: Halle on the Phoneme
3. AVERY ANDREWS, (morpho)phonemics; Halle's argument
4. , Russian i/y controversy
5. Harry Bochner, Re: Hades and Ulysses
6. "M. Sokolik", Re: (Morpho)phonology

Re: Hammond on English stress.

(1) I quite agree that the stress pattern of canasta vs. that of
orchestra is unpredictable, if that is what we mean by saying that
the former is due to exception marking. For, exception marking
is simply, in reality, a device for indicating that something is
unpredictable but not not quite admitting it. My own position
is that exception marking makes sense if there is factual evidence
for the special status of the exceptions. I know of two kinds
of evidence of this kind, but there may be others. One, the exceptional
forms form a closed, unproductive set. Two, the exceptional forms
are perceived as foreign. In the cases, we are discussing, I
think there is no such evidence, and hence to claim that English
stress is predictable is similar to saying that Mandarin has only
three tones (the fourth being due to exception marking), or that
there is no /s/ in English, it is really, say, /m/ but with
exception marking.

(2) Hades IS a personal name. And Ulysses, even though there is
nothing to flap in it, has the same (unstressed) last syllable that
Hades does, i.e., it rhymes with missies (at least, I think it does).

(3) I object to an underlying final /y/ instead of /I/ in industry,
Ogilvie, and so on (the SPE analysis) not because I object to
underlying representations in general (although I do) but because,
even assuming that URs are kosher, these particular ones are bizarre 
for the following reasons:

 (a) the resulting consonant clusters are highly unusual

 (b) there is no basis in alternations for such a contrast
 between /y/ and /I/, i.e., it is not true that words
 with /y/ resp. /I/ show up with these as phonetic values
 before some suffix (for example).

 (c) there is also no basis in analogy to other alternations.
 For example, it has been suggested (but I forget whether
 this is SPE or someone else, maybe McCawley??) that final
 -er can be derived from either /r/ or /Vr/, that this
 can be used to account for the stress pattern
 of mInister (cf. ministr-y, so underlyingly /mInIstr/,
 or carpenter (cf. carpentr-y). However, this won't work
 either because of examples like pilAster (cf. pilastr-ade).

Bruce Nevin's criticism of Halle's argument against the phoneme
misses the point, I believe. Halle quite correctly noted the
implausibility of any theory of phonology which had to claim
that the voicing in Russian [zhejhby] had to be due to a different
process than that in [mogby]. Simplicity metrics and such may
not have stood the test of time, but there ARE compelling reasons
for accepting this point anyway. For example, as Kiparsky pointed
out in 1968 (if not earlier), we do not find languages in which
the two parts of such a process (the allophonic and the neutralizing)
have a different diachrony. Likewise, I would add that synchronically
we do not seem to find languages in which automatic (natural, whatever)
processes crucially apply in just the allophonic or just the neutralizing
cases (and indeed this could not be the case if the diachronic generalization
is valid!). Schane, of course, had an argument to the contrary,
where phonemic status (or lack thereof) was a determinant of sound
change, but a recent article shows that all his own examples are
misanalyzed and that control cases to his claims show no effect
of phonemic status. I would be interested to know if anybody else
has any purported counterexamples to the Kiparsky generalization.

On the other hand, as I have pointed out both here and in print,
there is some evidence (first noted by Henryk Ulaszyn many years
before Halle) that native speakers perceive the allophonic effects
of a process differently from the neutralizing ones. Based on
this (correct) insight, Ulaszyn introduced precisely the phonemic
theory that Halle later attacked, again quite correctly, since all
the indications are that the PROCESS is the same. I am right, and you
are right, and everything's quite correct (as a better lyricist than I
once wrote)...

Some recent contributors to this newsletter have claimed to be
satisfied with Halle's argument against the autonomous phoneme (the one
reported in Chomsky (1966:88-89, Current Issues in Linguistic Theory, Mouton).

But I would claim that this argument rests on an incomplete understanding
of why arguments from simplicity are convincing, in those cases where they
actually are. The force of simplicity arguments derives from the fact
that significant generalizations need explanations, and, in many cases,
the most plausible explanation for a linguistic generalization is that
the various cases that it covers are all consequences of some single
facet of mental structure. E.g. `Det ... N ...' sequences have the
same internal structure in the various places where they occur because
these sequences are all organized by a common factor, which we used to
call the NP rule (now usually regarded as being an assortment of
parameter settings). Simplicity arguments usually seem compelling
in syntax, and in autosegmental & metrical phonology, but not in this case
of Halle's, basically because of the way in which historical development is
involved in the phenomenon.

What the argument does show is that people are capable of initiating
and propagating a sound change that sometimes has (observationally)
allophonic effects, sometimes phonemic ones. But the acquisitional
problem posed by a sound-change in progress is obviously quite different
from that posed by the alternations that the sound-change leaves behind
when it is completed, since only in the former case does the learner
have both the input and the output of the rule available in the data,
(whether the learner is a child learning it all at once, or an adult
acquiring only the change).

So it remains possible that the phonemic and morphophonemic branches
of the obstruent-voicing rule might be treated differently by learners,
once the rule has become obligatory & universal in the speech community.
For evidence that this isn't happening, one would need evidence of both
branches of the rule undergoing some further change in common, such as
loss or generalization.

Avery Andrews (ada612csc.anu.edu.au)

Alexis Manaster-Ramer brought up the question of the orthographic distinction
between 'i' and 'y' (yerih) in Russian. This distinction has led to a great
amount of debate among Russian linguists, because of Baudouin's original
analysis that the letters represent variant pronunciations of the same phoneme.
Alexis wrote:

>...Moreover, Baudouin himself
>was forced in later years to realize that there is something wrong,
>because (despite the rhyming facts!) his Polish and Russian colleagues
>etc. kept refusing to admit that 'i' and 'y' were intuitively the
>same sound. This also continues into the present: it is quite clear
>that Polish and Russian linguistics students have a lot of difficulty
>believing that these are the same sound. But, of course, the "psychophonetic"
>theory of the phoneme predicts that native speakers should not be able to
>distinguish the allophones of the same phoneme!

I did not know that Baudouin ever retreated an iota from his original position,
but his Moscow school descendants kept the controversy alive. It is worth 
noting that the issue is not so simple as Alexis states. All words beginning
with 'i' (high front vowel) are pronounced with 'y' (high back unrounded)
when preceded by nonpalatalized consonants. Thus, 'with Ivan' is pronounced
's yvanom' rather than 's ivanom' (although the spelling is still with 'i').
I believe that the ability of Russians to perceive the i/y distinction better
than other allophonic distinctions derives mainly from their having been taught
to spell--in just the same way that linguistic students can be taught to
perceive and write phonetic transcriptions. So the allophony of i/y may be
perceived differently because its orthographic representation is exceptional.
And it should also be pointed out that, while many native linguists disagreed
with Baudouin, many agreed (and still do). So Baudouin's admission that 
'something was wrong' was not necessarily an admission that his critics were
right.
 -Rick Wojcik (rwojcikatc.boeing.com)

David Pesetsky (pesetskATHENA.MIT.EDU) writes:
> Mike Hammond (HAMMONDccit.arizona.edu) writes:
> >If Hades is pronounced [heDiz] then we're committed to a similar
> >analysis: /hadi+s/. (Frankly, I have no clear intuition about whether
> >that word is plural or singular.)... 
> ... This suggests that Hades, Ulysses etc. are indeed
> pluralia tantum in English. Cf. "I need to buy three pants", which is
> worse than "three pairs of pants", but better than "three pantses".
> Thus, Hammond's suggestion might be right.

Consider:
(1) Hades _is_ (*are) the abode of the dead.
(2) Ulysses _is_ (*are) the hero of the Odyssey.
(3) The pants _are_ (*is) lying on the chair.

Thus I don't see how "Hades" and "pants" can be lumped together. Note
that presence of a plural marker is not the only possible cause of
infelicity of a plural form. "I know three Ulysses's." doesn't sound
great to me, but neither does it sound much worse than "I know three
Francis's.", and if I understand Hammond correctly, "Francis" cannot
contain /+s/ in his analysis because that would force the second vowel
to be tense.

More generally, this strikes me as the sort of "Diacritic use of
Juncture" that a constrained theory of Morphology would have to rule out.

-- Harry Bochner
-- bochnerdas.harvard.edu

Consider also 'kudos' [kuDos] listed recently in the Scrabble dictionary
as singular 'kudo'; I have seen this "singular form" used elsewhere in
print as well. In addition, just recently two students on two
separate occasions referred to Hade (apparently there are more than
one of them, thus Hades)--context made it clear that they weren't
talking about Haiti. I think we're onto something here.

[End Linguist List, Vol. 2, No. 140]